| About this Abstract |
| Meeting |
MS&T22: Materials Science & Technology
|
| Symposium
|
Manufacturing and Processing of Advanced Ceramic Materials
|
| Presentation Title |
High Throughput, Ultra-fast Laser Sintering of Alumina Sample Array for Establishing the Machine-learning-based Mapping Between Microstructure and Hardness |
| Author(s) |
Fei Peng, Hai Xiao, Dongsheng Li, Rajendra K Bordia, Jianhua Tong, Jianan Tang, Xiao Geng, Siddhartha Sarkar, Bridget Sheridan |
| On-Site Speaker (Planned) |
Fei Peng |
| Abstract Scope |
We report ultra-fast laser sintering of alumina that achieves the desired density and microstructure for alumina within ~10 seconds. Compared to furnace sintering, ultra-fast laser sintering can either suppress or enlarge the grain size for the same sintering density. A sample array of ~80 sample units (~500 μm × 500 μm × 100 μm each) can be sintered simultaneously under one laser scan, which results in various microstructures for each sample unit due to the laser power distribution. The hardness of each sample unit and corresponding microstructure were characterized to establish the datasets for machine learning (ML) training. The hardness vs. relative density data obtained from this high throughput method, well match the literature data. We developed ML algorithms that can precisely predict the laser-sintered alumina microstructure from the hardness values and also precisely predict the hardness of the laser-sintered alumina from the SEM micrographs with less than 5% error. |